Integration: directing work, knowledge and lessons learned

Integration is the project manager's distinctive job. Specialists can estimate costs, test kiosks or negotiate contracts; only the project manager is responsible for making all the pieces work together so that the project delivers its objectives. This lesson covers the execution side of integration: directing and managing the work, managing project knowledge and lessons learned, and monitoring and controlling project work by turning raw data into information and reports that drive decisions. Lesson 11 covers the other central integration activity, controlling changes.

What integration means

Project integration management includes the activities needed to identify, combine, unify and coordinate the various processes and activities of a project. In practice it means making trade-offs between competing objectives, keeping plans consistent with one another, managing dependencies between workstreams and between the project and its environment, and taking decisions with the whole project in view. When the electrician crew risk (R05) is mitigated by booking a second crew, integration is what makes sure the extra cost appears in the budget, the crew appears in the resource plan, the secondary risk of crowding is managed, and the schedule reflects the new dates.

Integration responsibility cannot be delegated. Team members and specialists manage their parts; the project manager remains accountable for the result as a whole. The PMBOK Guide's principles express related ideas: thinking in terms of systems and recognizing interactions, stewardship of resources, and tailoring the approach to the context. In the eighth edition's structure, much of what this lesson covers sits in the governance domain and runs across the executing and monitoring and controlling focus areas.

Directing and managing project work

Directing and managing project work means leading and performing the work defined in the project management plan and implementing approved changes, so that the project produces its deliverables. The project manager, with the team, carries out planned activities, acquires and manages resources, manages sellers through the procurement processes, implements risk responses, collects data on performance, and raises change requests when needed. Approved changes, including corrective actions, preventive actions and defect repairs, are also implemented here.

This activity produces deliverables and work performance data, the raw observations and measurements about the work: which kiosks were installed today, how many tags were encoded, hours worked, costs incurred, test results. It also produces entries in the issue log, change requests and updates to project documents. The project management information system (PMIS), whether a scheduling tool, a shared workspace, a ticketing system or a mix, is where this data is collected and shared.

Leadership is part of directing work. PMP-120 covered the people side: building the team, motivating, coaching and resolving conflict. The point here is that executing the plan is not mechanical. On Project Compass, during kiosk installation, the project manager spends more time unblocking problems, such as a missing network port, a delayed delivery or a staff member who needs training earlier than planned, than reading reports.

Monitoring and controlling project work

Monitoring and controlling project work means tracking, reviewing and reporting overall progress against the performance objectives in the plan. It turns work performance data into work performance information by comparing it with plans, and packages that information into work performance reports for decision makers. The three terms are easily confused, and the exam uses them.

# From work performance data to work performance information: kiosk installation week 31
data = [  # (kiosk, planned day, installed day or None, tests passed of 12)
    ("K1", 1, 1, 12), ("K2", 1, 2, 11), ("K3", 2, 2, 12), ("K4", 2, 4, 12),
    ("K5", 3, None, 0), ("K6", 3, 3, 9),
]
done = [d for d in data if d[2] is not None]
late = [d[0] for d in done if d[2] > d[1]]
not_started = [d[0] for d in data if d[2] is None]
first_pass = [d[0] for d in done if d[3] == 12]
print(f"Installed {len(done)} of {len(data)} planned by day 4; late: {', '.join(late)}; "
      f"not installed: {', '.join(not_started)}")
print(f"First-pass test yield {len(first_pass)}/{len(done)} = {len(first_pass) / len(done):.0%} "
      f"(target 90 %)")
fails = sum(12 - d[3] for d in done)
print(f"Failed test steps: {fails}, of which {12 - 9} on K6 (card reader) and 1 on K2 (receipt printer)")

Installed 5 of 6 planned by day 4; late: K2, K4; not installed: K5
First-pass test yield 3/5 = 60% (target 90 %)
Failed test steps: 4, of which 3 on K6 (card reader) and 1 on K2 (receipt printer)

The list of kiosks with planned and actual days and test results is work performance data. The statements that two kiosks were installed late, one is not yet installed and the first-pass test yield is 60 % against a 90 % target are work performance information: they compare data with the plan and the quality metric. The weekly status report to the sponsor, which summarizes the kiosk position alongside the schedule, cost and risk picture and recommends action, is a work performance report.

Monitoring also covers forecasting (the EAC and finish-date forecasts from PMP-130), variance analysis, trend analysis, root cause analysis and the review of risks and issues. Its outputs include change requests: corrective actions to bring performance back in line with the plan, preventive actions to reduce the chance of future problems, and defect repairs. Change requests then go to integrated change control.

Managing project knowledge

Managing project knowledge means using existing knowledge to achieve project objectives and creating new knowledge that contributes to organizational learning. Knowledge comes in two forms. Explicit knowledge can be written down and shared through documents, databases and registers: procedures, specifications, lessons learned. Tacit knowledge is personal and hard to express: the experience, judgment and know-how that lets an experienced fit-out lead see a problem coming. Tacit knowledge is shared mainly through people working together: mentoring, pairing, communities of practice, workshops and storytelling.

A lessons learned register records challenges, problems, risks and opportunities encountered, and what worked or did not, so that the project and future projects can benefit. Good practice is to capture lessons throughout the project, not only at the end, when memories have faded and people have moved on, and to write them so they can be found and used: specific, actionable and tagged by topic.

# Querying the library's lessons learned register before planning the next stage
register = [
 {"id": "LL-014", "project": "Westfield branch refit", "area": "procurement",
  "lesson": "Tag supplier lead time doubled in December; order by end of October", "tags": {"rfid", "lead time"}},
 {"id": "LL-022", "project": "Central library kiosks", "area": "quality",
  "lesson": "Soak-test kiosks for two weeks with real patrons before opening", "tags": {"kiosk", "testing"}},
 {"id": "LL-031", "project": "Westfield branch refit", "area": "risk",
  "lesson": "Mall landlords restrict noisy work to 7-9 am; confirm in the lease schedule", "tags": {"landlord", "schedule"}},
 {"id": "LL-040", "project": "Library system migration", "area": "integration",
  "lesson": "Freeze the item data mapping before the vendor starts encoding", "tags": {"rfid", "data"}},
 {"id": "LL-047", "project": "Central library kiosks", "area": "procurement",
  "lesson": "Specify card-reader firmware version and update rights in the contract", "tags": {"kiosk", "contract"}},
]
def search(*words):
    hits = [r for r in register if any(w in r["tags"] or w in r["lesson"].lower() for w in words)]
    return hits
for query in (("kiosk",), ("rfid", "tag")):
    print("Query:", " or ".join(query))
    for r in search(*query):
        print(f"  {r['id']} [{r['area']}] {r['lesson']}")

Query: kiosk
  LL-022 [quality] Soak-test kiosks for two weeks with real patrons before opening
  LL-047 [procurement] Specify card-reader firmware version and update rights in the contract
Query: rfid or tag
  LL-014 [procurement] Tag supplier lead time doubled in December; order by end of October
  LL-040 [integration] Freeze the item data mapping before the vendor starts encoding

Before kiosk installation began, the technology lead searched the library's register. Two entries from the central library's earlier kiosk project stood out: soak-test kiosks for two weeks with real patrons before opening, and specify the card-reader firmware version and update rights in the contract. The second had already shaped the Eastgate contract (lesson 9). The first had not been acted on, because lesson 5 chose a contract clause rather than a soak test to deal with R07. The day 4 data, with K6's card reader failing three test steps, made the case: the technology lead raised a preventive change request for a two-week soak test costing $5,000, which became CR-06 in lesson 11. Using knowledge from earlier projects is cheaper than relearning it.

Knowledge management depends on culture as much as tools. People share lessons, especially mistakes, only if they trust that doing so will not be held against them. Retrospectives in agile teams, after-action reviews and blameless post-incident reviews all depend on psychological safety, which PMP-120 discussed. A project manager who reacts to bad news by looking for someone to blame will soon stop hearing it.

Integration in adaptive and hybrid projects

In adaptive projects, much integration is done by the team: the product owner integrates stakeholder needs into a prioritized backlog, and the team integrates work daily through practices such as continuous integration. The project manager or delivery lead focuses on the environment around the team, removing impediments, coordinating with other teams and stakeholders, and making sure that governance requirements are met without slowing the team down.

In a hybrid project, integration includes joining the predictive and adaptive parts. For Project Compass, the integration points are the soft opening, when the app's first release must work with the kiosks, the library system and the branch's network, and the grand opening. The project manager keeps an integration plan that lists dependencies between the fit-out, technology and app workstreams, holds a fortnightly integration meeting with each workstream lead, and reports one status to the board that combines earned value for the predictive parts with release burn-up for the app.